A method for detection and/or tracking of objects in motion 16 in a scene under surveillance 15, in which besides the objects in motion 16, interfering objects and/or interfering regions—both hereinafter called interfering factors 17, 23—can occur, is proposed, in which the scene under surveillance 15, a plurality of regions are defined that are divided up into various region classes; and a first region class D1, 5 includes sensitive regions, in which no and/or only insignificant interfering factors 17, 23 are located and/or are to be expected; and for detection and/or tracking of the objects in motion 16 in the sensitive regions, a sensitive content analysis is performed, and a second region class D2, 6 includes semi-sensitive regions 19, in which interfering factors 17, 23 are located and/or are to be expected, and for detection and/or tracking of the objects in motion 16 in the semi-sensitive regions 19, a semi-sensitive content analysis is performed, which is limited and/or modified compared to the sensitive content analysis.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A method for detection and tracking of objects in motion ( 16 ) in a scene under surveillance ( 15 ), in which scene, besides the objects in motion ( 16 ), interfering factors ( 17 , 23 ) comprising interfering objects, interfering regions, or both are included, comprising steps of: defining the scene under surveillance ( 15 ) as a plurality of regions that are defined as image sections in the scene under surveillance ( 15 ) that are stationary over the surveillance period, for each frame in an image sequence of frames derived from the scene under surveillance; dividing up the plurality of regions into various region classes, the various region classes including: a first region class (D 1 , 5 ) with sensitive regions in which no interfering factors ( 17 , 23 ) or only insignificant interfering factors ( 17 , 23 ) are located, are expected or are located and expected, a second region class (D 2 , 6 ) with semi-sensitive regions ( 19 ), in which interfering factors ( 17 , 23 ) are located, are expected, or are located and expected; and a third region class (D 3 ) with insensitive regions in which interfering factors ( 17 , 23 ) comprising blind spots are located; performing a sensitive content analysis for detection, for tracking, or both of the objects in motion ( 16 ) in the first region class (D 1 , 5 ) with sensitive regions; and performing a semi-sensitive content analysis for detection, for tracking, or both of the objects in motion ( 16 ) in the second region class (D 2 , 6 ) with semi-sensitive regions ( 19 ), which semi-sensitive content analysis is limited, is modified or both, as compared to the sensitive content analysis; wherein no content analysis for detection, for tracking or both of the objects in motion ( 16 ) is performed in the third region class (D 3 ) with insensitive regions.
2. The method as defined by claim 1 , wherein in one of the semi-sensitive regions ( 19 ), objects in motion ( 16 ) are detected, are tracked or are detected and tracked that exhibit an unconventional pattern of motion compared to the semi-sensitive regions ( 19 ).
3. The method as defined by claim 2 , wherein in one of the semi-sensitive regions ( 19 ), objects in motion ( 16 ) are detected, are tracked or are detected and tracked that move in the semi-sensitive regions ( 19 ) counter to a general direction of motion.
4. The method as defined by claim 2 , wherein in one of the semi-sensitive regions ( 19 ), objects in motion ( 16 ) are detected, are tracked or are detected and tracked that in the semi-sensitive regions ( 19 ) move with a directional motion in front of a scene area ( 15 ) that is non-moving.
5. The method as defined by claim 2 , wherein in the semi-sensitive regions ( 19 ), objects in motion ( 16 ) are detected, are tracked or are detected and tracked by way of analysis of the optical flow.
6. The method as defined by claim 1 , wherein in the semi-sensitive regions ( 19 ), objects in motion ( 16 ) are detected, are tracked or are detected and tracked via a comparison of characteristics of the objects in motion ( 16 ) with characteristics of reference objects.
7. The method as defined by claim 6 , wherein the reference objects or their characteristics are stored in memory in a reference database ( 10 ).
8. The method as defined by claim 1 , wherein in the semi-sensitive regions ( 19 ), objects in motion ( 16 ) are detected, are tracked, are detected and tracked or are further tracked on the basis of any combination of information from the group consisting of motion information, object information, and characteristic information, which information is detected in the sensitive regions ( 20 ).
9. An apparatus ( 1 ) for performing a method for detection and tracking one or more objects in motion ( 16 ) in a scene under surveillance ( 15 ), comprising: an interface ( 3 ) configured for receiving image sequences from the scene under surveillance, in which scene under surveillance, besides the one or more objects in motion ( 16 ), interfering factors ( 17 , 23 ) comprising interfering objects, interfering regions, or both are included, a classification module ( 4 ) configured for defining regions in the scene under surveillance defining the scene under surveillance ( 15 ) as a plurality of regions, which regions are defined as image sections in the scene that are stationary over a surveillance period, for each frame in an image sequence of frames derived from the scene and dividing up the regions into various region classes, for each image sequence frame received from the scene under surveillance, wherein a first region class includes sensitive regions ( 20 ), within which no significant interfering factors ( 17 , 23 ) or only insignificant interfering factors ( 17 , 23 ) are located, are expected or are located and expected, and a second region class includes semi-sensitive regions ( 19 ) in which interfering factors ( 17 , 23 ) are located, are expected or are located and expected and a third region class includes interfering factors comprising blind spots; a first analysis module ( 5 ) configured for performing a sensitive content analysis for detection, tracking or detection and tracking of the objects in motion ( 16 ) in the sensitive regions ( 20 ); and a second analysis module ( 6 ) configured for performing a semi-sensitive content analysis, for detection, for tracking or for detection and tracking of the objects in motion ( 16 ) in the semi-sensitive regions ( 19 ), which analysis is limited, is modified or both limited and modified compared to the sensitive content analysis; wherein the first analysis module ( 5 ) and the second analysis module ( 6 ) perform no content analysis for detection, for tracking or both of the objects in motion ( 16 ) in the third region class with insensitive regions.
10. A computer program having program code means, which program code means is stored in a non-transitory computer-readable medium and performs all the steps of the method for detection and tracking of objects in motion in a scene under surveillance, as set forth in claim 1 , when the computer program is executed on a computer or an apparatus.
11. The method for detection and tracking of objects in motion ( 16 ) in a scene under surveillance ( 15 ), as set forth in claim 1 , wherein the step of dividing up the plurality of regions into various region classes is performed manually by a user or in an automated fashion using an image-based, content analysis of the scene under surveillance ( 15 ).
12. The method for detection and tracking of objects in motion ( 16 ) in a scene under surveillance ( 15 ), as set forth in claim 1 , wherein the step of sensitive content analysis comprises video content analysis.
13. The method for detection and tracking of objects in motion ( 16 ) in a scene under surveillance ( 15 ), as set forth in claim 12 , wherein the video content analysis includes forming or taking on a scene reference image, segmentation of objects, and detection and/or tracking of the segmented objects over time.
14. The apparatus as set forth in claim 9 , wherein the classification module ( 4 ) divides up the plurality of regions into various region classes manually by a user or in an automated fashion using an image-based, content analysis of the scene under surveillance ( 15 ).
15. The apparatus as set forth in claim 9 , wherein the first analysis module ( 5 ) implements video content analysis as the sensitive content analysis.
16. The apparatus as set forth in claim 15 , wherein the video content analysis includes forming or taking on a scene reference image, segmentation of objects, and detection and/or tracking of the segmented objects over time.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
September 2, 2008
November 4, 2014
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.